In this first clinical trial femoral vein transposition (FVT) was prospectively compared with its "graft" counterpart. Patients requiring vascular access due to occluded central veins were divided into two groups: group I (n = 17) underwent femoral loop grafts and group II (n = 15), FVT. Complications were recorded prospectively and patency rates were compared by the Kaplan Meier method with statistical comparisons by Mann Whitney U and chi-square tests as appropriate. Between 1999 and 2003 30 patients underwent 32 operations. Mean follow-up was 8.9 months in group I and 6.8 months in group II (P >.05). Three group I patients developed a steal, requiring closure of the fistula in one. Five group II patients developed a steal, requiring closure of the fistula in one. The incidence of steal was similar in both groups (P >.05). The infection rates were similar: 35.3% and 26.7% in group I versus group II (P >.05), respectively. Among the six infections in group I, four patients required excision of the graft due to infection, whereas only one arteriovenous fistula (AVF) was closed in group II for infection. The 1-year primary patency rate in group I was 37.5% versus 86.7% in group II (P <.05). In conclusion, Femoral AVF grafts have been condemned because of the unacceptable high rate of septic complications. FVT obviates the need for a graft and has fewer infectious problems and better patency rates.